OXYGEN GENERATION THROUGH ELECTRICALLY STIMULATED O2 REDUCTION

ACROSS ION EXCHANGE MEMBRANES

A Major Qualifying Project Report

Submitted to the Faculty of

WORCESTER POLYTECHNIC INSTITUTE

In partial fulfillment of the requirements

for the Degree of Bachelor of Science

by

_________________________

Devin Churchman, CM

_________________________

Daniel Jones, CH

_________________________

Raj Patel, CM

April 30, 2014

Approved:

_________________________________

Professor Ravindra Datta, Major Advisor

_________________________________

Professor Stephen J. Kmiotek, Co-Advisor

_________________________________

Professor Drew Broduer, Co-Advisor

2

Table of Contents List of Tables ................................................................................................................................................. 4

List of Figures ................................................................................................................................................ 5

Abstract ........................................................................................................................................................ 6

1. Introduction .............................................................................................................................................. 7

2. Literature Review ................................................................................................................................... 10

2.1 Social Implications............................................................................................................................. 10

2.1.1 Diseases and Conditions ............................................................................................................ 10

2.1.2 Current Solutions ....................................................................................................................... 11

2.2 Chemistry .......................................................................................................................................... 13

2.2.1 Four Electron Oxygen Reduction Reactions ............................................................................... 13

2.2.2 Two Electron Oxygen Reduction Reactions ............................................................................... 14

2.3 Proton Exchange Membrane (PEM) ................................................................................................. 18

2.3.1 Introduction to PEMs ................................................................................................................. 18

2.3.2 Mechanism ................................................................................................................................. 19

2.3.3 Hydrogen Oxygen Fuel Cell ........................................................................................................ 23

2.3.4 Oxygen generation case studies ................................................................................................ 25

2.4 Anion-Exchange Membrane (AEM)................................................................................................... 33

2.4.1 Development .............................................................................................................................. 33

2.4.2 Mechanism ................................................................................................................................. 34

2.4.3 Benefits and Challenges ............................................................................................................. 36

2.4.4 Oxygen Generation Case Studies ............................................................................................... 37

3. Methodology .......................................................................................................................................... 38

3.1 Conceptual Design ............................................................................................................................ 38

3.2 Apparatus .......................................................................................................................................... 40

3.3 Materials ........................................................................................................................................... 42

3.4 Experimental ..................................................................................................................................... 43

4. Results & Discussion ............................................................................................................................... 47

4.1 Liquid Electrolysis .............................................................................................................................. 47

4.2 Vapor Electrolysis .............................................................................................................................. 48

4.3 Two Electron Oxygen Reduction ....................................................................................................... 52

4.4 Carbon Degradation .......................................................................................................................... 55

3

5. Conclusion and Recommendations ....................................................................................................... 57

5.1 Conclusion ......................................................................................................................................... 57

5.2 Recommendations ............................................................................................................................ 58

5.2.1 Use of Alternative Catalysts ....................................................................................................... 58

5.2.2 Fabricating Membranes ............................................................................................................. 60

5.2.3 Stacking ...................................................................................................................................... 61

5.2.4 Anion Exchange Membranes ..................................................................................................... 62

5.2.5 Mathematical Analysis ............................................................................................................... 62

Works Cited ................................................................................................................................................ 64

Appendix A: Results summary table .......................................................................................................... 68

Appendix B: MEA P1 Results ...................................................................................................................... 76

IR Spectroscopy of MEA P1 ................................................................................................................. 77

Appendix C: U.S Patent 5,211,984 for Membrane Catalyst Loading in MEA Fabrication (Wilson, 1993)78

Appendix D: Sample plot generated by COMSOL. .................................................................................... 82

4

List of Tables Table 2.1 A The Half Reaction and Overall Reaction for MA's O2 Concentrator (Ma & Yu, 1995) ............. 12

Table 2.2 A The Four Electron Oxygen Reduction Reactions ...................................................................... 13

Table 2.2 B Water Electrolysis .................................................................................................................... 14

Table 2.2 C The Two Step Production of Hydrogen Peroxide using Anthraquinone .................................. 15

Table 2.2 D The Two Electron Oxygen Reduction Reactions ...................................................................... 15

Table 3.3 A Summary of Tested MEAs ........................................................................................................ 43

Table 4.1 A MEA I2 Liquid Electrolysis Results ............................................................................................ 47

Table 4.2 A Vapor Electrolysis Results at Ambient Temperature ............................................................... 49

Table 4.2 B Vapor Electrolysis Results at Elevated Temperature ............................................................... 50

Table 4.3 A Test result for MEA P3 with He bubbled through H2O2 fed to the cathode and humidified air

fed to the anode ......................................................................................................................................... 54

Table 4.3 B Test results for MEA I2 with humidified air feed to the cathode and He bubbled through

H2O2 feed to the anode ............................................................................................................................. 55

Table 4.4 A Surface Carbon Degradation .................................................................................................... 56

5

List of Figures Figure 2.2 A Models for the Substitution of Carbon by Nitrogen Atoms at the Edges of the Carbon Sheets (Boehm et al., 1984) .................................................................................................................................... 18 Figure 2.3 A Sulfuric Acid Analog of Polymer Electrolyte (Sulfuric Acid Model)......................................... 19

Figure 2.3 B Chemical Structure of Nafion® (Zhou, et al 2007) .................................................................. 21

Figure 2.3 C Reverse micellar Cluster-Network Structure of Hydrated Nafion® ........................................ 22

Figure 2.3 D Proton Diffusion via (a) en masse or vehicle and (b) Grotthuss Mechanism. ........................ 23

Figure 2.3 E Sample Proton Exchange Membrane Fuel Cell Schematic (Lister & McLean, 2004) .............. 24

Figure 2.3 F Graphical representation for the oxygen concentration over time (Fujita, Nakamura & Muto,

1985) ........................................................................................................................................................... 26

Figure 2.3 G Change in voltage of the PEM Fuel Cell over time (Fujita, Nakamura & Muto, 1985). .......... 27

Figure 2.3 H Schematic view of the 25 cm2 cell for oxygen extraction (Eladeb et al., 2012) ..................... 29

Figure 2.3 I Current Density vs. Air Flow (Eladeb et al., 2012). ................................................................... 30

Figure 2.3 J Current Density vs. Cell Voltage (Eladeb et al., 2012) ............................................................. 31

Figure 2.3 K Current Density vs. Time (Eladeb et al., 2012) ........................................................................ 31

Figure 2.3 L O2 Outlet Flow vs. the Inlet O2 Flow over Current (Eladeb et al., 2012) ................................ 32

Figure 2.3 M Current Density vs. the ORR efficiency (Eladeb et al., 2012) ................................................. 33 Figure 2.4 A Schematic of dissociation and solvation of the pendant OH- groups within the pores of a

hydrated AEM (Grew et al., 2010) .............................................................................................................. 35 Figure 3.1 A Electrolysis Aided PEM Pump Model ...................................................................................... 38

Figure 3.1 B Two Electron Oxygen Reduction Reaction PEM Pump ........................................................... 39 Figure 3.2 A Example Schematic of Experimental Procedure ..................................................................... 40

Figure 3.2 B Diagram for the fuel cell where the cathode is receiving the humidified inlet while the anode

is receiving a dry inlet. In addition, the fuel cell is connected to the power supply and the Fuel Cell Test

System ......................................................................................................................................................... 40

Figure 3.2 C Overall Diagram of the Fuel Cell Test bed located in the Fuel Cell Center at Worcester

Polytechnic Institute ................................................................................................................................... 41

Figure 3.2 D HP power supply used for the experimental runs .................................................................. 42 Figure 6.1 A Overview of different electro catalysts for H2O2 production (Siahrostami et al., 2013) ...... 59 Figure 6.3 A Sample Fuel Cell Stack (Fuel Cell Store, 2013) ........................................................................ 61 Figure 6.4 A Sample Anion Exchange Membrane for metal cation-free alkaline fuel cell ......................... 62 Figure 6.5 A Geometry of a proton exchange membrane (PEM) modeled in COMSOL Multiphysics.

(COMSOL Multiphysics, 2014) .................................................................................................................... 63

6

Abstract Oxygen transport through a proton exchange membrane (PEM) fuel cell was examined

with the ultimate goal of creating a model for a portable oxygen generator. Water electrolysis by

four electron oxygen reduction (ORR) along with two electron ORR were tested using membrane

electrode assemblies (MEAs) with Pt/C, carbon (Printex L6), and PtIrB catalysts. Results in trials

for all configurations yielded small currents and little to no oxygen production. Based on this

study, it appears a vapor electrolysis PEM fuel cell oxygen pump and two electron oxygen

reduction based on Pt/C and Printex L6 are not feasible. Alternative catalysts for two electron

oxygen reduction on a PEM and alternative membranes may lead to more functional models of

the proposed electrochemical oxygen generator.

7

1. Introduction There are numerous diseases and conditions that affect the respiratory system. While

these conditions vary in severity, they all affect those living with these conditions in their

everyday lives. Current solutions to respiratory problems, from portable oxygen tanks to oxygen

concentrators, can be large, heavy, inconvenient, and can cause potential safety hazards. A new

design for a personal and portable device was based on PEM fuel cell. Membrane electrode

assemblies (MEAs) would be used as a means to output an enriched oxygen stream that could be

delivered directly to the person.

The process is based on the oxygen reduction reaction (ORR). ORR would occur at the

cathode while oxygen evolution would occur at the cathode, meaning the anode exhaust would

be the oxygen enriched stream. There are two types of ORR, the four electron reaction and the

two electron reaction. The four electron ORR reacts via water electrolysis, while the two electron

ORR uses hydrogen peroxide as an intermediate. The two electron ORR is preferable to the four

electron due to the lower energy costs associated with the reaction.

When considering the MEA, there are two possible ion exchange membranes that could

be used in this design: the proton exchange membrane (PEM) and the anion exchange membrane

(AEM). The key difference between the two membranes is that the PEM facilitates the exchange

of protons across the membrane while the AEM facilitates the exchange of anions across the

membrane. Both membranes have advantages to their respective use in this design, however

given the established research on PEMs and the relative infancy of AEMs, the PEM was chosen

for the MEAs to be used in this study.

8

Oxygen extraction has been accomplished using PEM electrolysis in some studies (Eladeb

et al., 2012), although this study was performed using liquid electrolysis as opposed to vapor

electrolysis, which is the ultimate goal of this design. It is important to compare these liquid

electrolysis results with the gathered vapor electrolysis results to assess the validity of the vapor

PEM electrolysis and its viability as an oxygen pump.

Experimental trials consisted of varying feeds, temperature conditions, and applied

voltages to the MEA being tested. A schematic of the experimental system can be seen in Figure

3.2A. Three sets of Nafion® 115 membranes were used with differing catalysts: Pt/C catalyst at

both the cathode and the anode, Printex L6 carbon catalyst at the cathode and Pt/C at the anode,

and Pt/C at the cathode and unsupported PtIrB at the anode. The voltage was set given the type

of ORR being pursued in the trial, voltages lower than 1.2 V being for the two electron ORR while

higher voltages were intended to induce the four electron ORR.

The experimental results proved that the initial goal of designing an oxygen generator had

been unsuccessful. Neither the Pt/C catalyst nor the Printex L6 catalyst was successful in fostering

the two electron ORR in the system. The MEAs loaded with the PtIrB, which were intended for

the vapor electrolysis trials, also failed to yield promising results. With little to no observable

oxygen evolution at the anode and very small sustainable current through the cell, it would

appear that vapor electrolysis is not an efficient method for the transport of oxygen through

PEM.

At the conclusion of this study, it was determined that none of the proposed designs as

tested in these experiments would yield any kind of practical and efficient oxygen generator

9

device. Pt/C and Printex L6 catalysts failed to facilitate any two electron ORR across the

membrane. Any attempt to further the two electron ORR study using a PEM must be done with

alternative catalysts that show more activity for the two electron reaction. Vapor electrolysis did

not yield promising enough results to warrant further examination into this method. The low

current densities observed along with the mass transport limitations of the system show that the

scale up of this system is not worth pursuing.

It is recommended that any further study on this design focus on promoting the two

electron ORR for oxygen transport across the PEM. There is the potential for other catalysts to

be more active for the two electron ORR, which could potentially lead to a more practical design

and a usable model. Oxygen transport using AEM may possibly yield more favorable results,

however more research must first be performed on the subject. COMSOL Multiphysics, a physics

modeling software, could potentially be used by researchers to perform theoretical calculations

and model the MEA before performing future experiments.

10

2. Literature Review

2.1 Social Implications There are a number of activities in society that people perform for personal and

professional reasons that have a negative effect on the respiratory system when performed

repeatedly over long periods of time. The most common way people damage their respiratory

system is smoking tobacco. Additionally, there are several materials used in industry that can

cause lung damage to those who work with the raw materials and those who use the final

product. All of these everyday activities can lead to numerous respiratory diseases and

conditions. Finally, air pollution causes respiratory ailments.

2.1.1 Diseases and Conditions

There are a variety of diseases and conditions that affect the respiratory system.

According to the UCSF Medical Center, these ailments are divided into four categories; 1)

occupational lung diseases which are caused by long term inhalation of industrial irritants such

as beryllium, silica, and asbestos; 2) chronic obstructive pulmonary disease (COPD), which is

primarily caused by years of tobacco smoking and includes the disease emphysema, which is the

fourth leading cause of death in the USA; 3) non-tuberculosis mycobacteria (NTM) which is

caused by a group of bacteria normally found in soil and water, and 4) interstitial lung disease

(ILD) the causes of which are mostly unknown (UCSF Medical Center, 2002). All of these ailments

damage the lung so that it cannot absorb the required amount of oxygen from the air into the

blood stream. As such, part of the treatment for these diseases is oxygen therapy, which is quite

simply to provide higher concentrations of oxygen to the patients so that their lungs can absorb

the necessary amount of oxygen. Oxygen therapy is tailored to the individuals exact condition so

that the amount of oxygen supplied varies from cases to case and can be anything from 30% to

11

98% oxygen. Duration can be for short term use in some cases such as lung infections in which

the lungs will generally recover, but the majority of patients on oxygen treatment are on it for

the rest of their lives (UCSF Medical Center, 2002).

2.1.2 Current Solutions

There are 3 major ways in which oxygen is stored or generated for such medical use. In

hospitals, where the demand for oxygen is high, it is stored in liquid form in chilled tanks. In

smaller medical facilities and for home use oxygen is stored in compressed gas cylinders. The

large oxygen cylinders can hold 6,500 standard liters of oxygen which will last about 2 days and

the smaller portable oxygen cylinders hold 164 or 170 liters and last four to six hours. The last

method is to generate oxygen with a personal oxygen concentrator which eliminates the need

for storage and regular deliveries of bulk oxygen cylinders. Personal oxygen concentrators for

medical purposes most commonly produce oxygen by removing nitrogen from the air via

nitrogen adsorption. In this process there are two steps; first nitrogen is adsorbed onto a packed

zeolite bed at high pressure, providing an enriched oxygen air exhaust. The second step is to

purge the bed of nitrogen by dropping the pressure to below atmospheric pressure (Gauthier,

Hendricks & Babcock, 1980). The currently available personal portable oxygen concentrators

work on this principle. They are priced around $3,000 – $4,000 and are generally the size of a

large laptop bag or small backpack and they can be used in portable application. Most of them

have the option of providing ether a continuous flow of O2 enriched air or else give a periodic

pulse of pure oxygen. However, they only have an average battery life of 2.5 – 3 hours, with a

few models offering extended battery life at the expense of the weight and size.

12

In 1995, Ma and Yu (1995) published a paper on a novel electrochemical oxygen

concentrator designed for medium scale use in less developed areas. The device produced 36 L

of 99.5% pure O2 an hour. This was achieved through a 2 electron Oxygen Reduction Reaction

(ORR) mechanism resulting in oxygen being pumped from the cathode to the anode. The overall

reactions at the electrodes and the complete cell are provided in Table 2.1A. In the first reaction

air and electricity are fed to a carbon cathode where O2 is reduced to a hydro peroxide ion. Next

the hydro peroxide ion is decomposed on a Mn(NO3)2 6H2O mesh to produce oxygen and

hydroxide. The resulting hydroxide is then transferred through the 7 M KOH electrolyte to a nickel

mesh anode where it is oxidized to produce oxygen (Ma & Yu, 1995).

Table 2.1 A The Half Reaction and Overall Reaction for MA's O2 Concentrator (Ma & Yu, 1995)

Cathode Reaction O2 (air) + H2O + 2e- → HO2- + OH- Eq 1

Mn(NO3)2 6H2O mesh Reaction HO2- → ½ O2 + OH- Eq 2

Anode Reaction 2 OH- → ½ O2 + H2O + 2e- Eq 3

Overall Reaction O2 (cathode) → O2 (Anode) Eq 4

This provides an attractive alternative to other forms of oxygen production. When

compared to the traditional method of oxygen generation, water electrolysis, there are many

advantages. The first of which is a lower power requirement; Ma’s device uses a 2 electron ORR

which theoretically only requires 0.48V along with 2 e- consumed by the O2 being pumped,

whereas water electrolysis requires higher than 1.23V. Despite the lower energy consumption,

the O2 production remains equivalent to that in water electrolysis. Additionally this method does

13

not consume water. This system has reportedly been used in several hospitals in China with no

reduction in performance after 12 months of continuous use (Ma & Yu, 1995).

2.2 Chemistry The ORR is a reaction in which O2 gains electrons (e-). Oxygen reduction in aqueous

solution generally proceeds by one of two routes, a four e- or a two e- pathway. These reactions

are already widely used in electrochemistry for power generation and H2O2 production.

2.2.1 Four Electron Oxygen Reduction Reactions

The 4 e- ORR is most commonly used for power generation, via a fuel cell. The overall

reactions for four e- ORR in acidic and alkaline electrolyte can be seen in Table 2.2A. These

reactions are not spontaneous, comprise several steps, and generally require a catalyst. Many

different catalysts have been used and even more are being researched. Among these current

and developing catalysts are noble metals, carbon materials, quinone and derivatives, and

several transition metal compounds (Song & Zhang, 2008). However, the most common catalyst

used commercially today is Pt supported on carbon.

Table 2.2 A The Four Electron Oxygen Reduction Reactions

In Acidic aqueous solution O2 + 4H+ + 4e- → H2O 1.229 V

In Alkaline aqueous solution O2 + H2O + 4e- → 4OH 0.401 V

The reverse reaction of the 4 e- ORR is involved in water electrolysis, as seen in Table 2.2B.

Water electrolysis is the simple process of running a sufficient current through water to produce

hydrogen and oxygen. This is important to note for the purposes of our paper as this sets the

upper limit of our own study. Additionally, it tells us that if we proceed through the four e- ORR

14

we will produce hydrogen as well as oxygen, a rather undesirable outcome due to the explosive

nature of hydrogen.

Table 2.2 B Water Electrolysis

2 H2O → 2 H2 + O2 1.229 V

2.2.2 Two Electron Oxygen Reduction Reactions

Hydrogen peroxide is very important in industry as an effective and clean way to purify

waste water. For this reason, ways of improving hydrogen peroxide production are constantly

under study. Currently there are two primary methods of hydrogen peroxide generation. The

older method is the electrolysis of aqueous solutions of H2SO4, KHSO4, or NH4HSO4. More

commonly hydrogen peroxide is produced through the hydrogenation, reduction, and then

oxidation of anthraquinone derivatives, as seen in Table 2.2C. This is an efficient production

method as only hydrogen, atmospheric oxygen, and water are consumed (D. Considine (Ed.),

1974). However, an electrochemical method (Assumpcao et al, 2012) would have two primary

advantages over the anthraquinone method. Mainly that it would take less energy to acquire

protons from an acidic solution rather than generate H2 separately. Additionally it is believed that

an electrochemical method would be able to achieve a higher efficiency than the 90% conversion

rate that the anthraquinone method yields (Panizza & Cerisola, 2008).

15

Table 2.2 C The Two Step Production of Hydrogen Peroxide using Anthraquinone

Step 1 C6H4:(CO)2:C6H3R + H2 → C6H4:(COH)2:C6H3R

Step 2 C6H4:(COH)2:C6H3R + O2 → C6H4:(CO)2:C6H3R + H2O2

More recently the electrochemical generation of hydrogen peroxide has received greater

attention. The electrochemical generation of H2O2 utilizes the 2 e- ORR as can be seen in Table

2.2D. To produce H2O2 one needs a catalyst that will only reduce oxygen partially, otherwise the

H2O2 will be further reduced to water. Carbon is believed to be one of the best catalysts for this

reaction due to its large surface area, corrosion resistance, and low price. However carbon has

many forms which possess a large range of varying properties (Sudoh, Kitaguchi, & Koide, 1985).

Table 2.2 D The Two Electron Oxygen Reduction Reactions

In Acidic aqueous solution O2 + 2H+ + 2e- → H2O2 0.70 V

H2O2 + 2H+ + 2e- → 2 H2O 1.76 V

In Alkaline aqueous solution O2 + H2O + 2e- → HO2 - + OH- –0.065 V

HO2 - + H2O + 2e- → 3OH- 0.867 V

Soltani et al. (2012) have explored the idea of generating hydrogen peroxide in situ by

electrochemical means. For their experiment they used an undivided cell with a Pt anode and a

gas diffusion cathode (GDC), through which they feed oxygen at a rate of 1 L/min. They tested

several forms of carbon catalysts coated onto the GDC; including carbon black(CB) –PTFE,

powdered activated carbon(PAC) –PTFE, carbon nano tube(CNT)-PTFE, and CB-CNT-PTFE. After a

run time of 40 minute with an applied current of 200 mA they found that the CB-CNT-PTFE (123.5

16

µM) did the best and the PAC-PTFE (58.45 µM) did the worst. However, because the CB-PTFE

(112.3 µM) and CNT-PTFE (100.9 µM) preformed similarly and costs significantly less than the CB-

CNT-PTFE, they used the CB-PTFE as their catalyst for all subsequent experiments (Soltani et al.,

2012).

In addition to testing for an effective catalyst, they also tested the effects of initial pH,

electrolyte concentration, and applied current on the generation of hydrogen peroxide. They

tested initial pHs between pH 2 – pH 9 and found the best conditions were at pH 3 and above pH

7. Above pH 7 the hydrogen peroxide primarily exists as the hydroperoxide ion, which is stable in

basic solution. On the other hand, acidic solutions tend to reduce the hydrogen peroxide to water

and the catalyst will facilitate the formation of H2 from acidic protons in the solution. Next these

investigators examined H2O2 generation at different electrolyte concentrations. They used

Na2SO4 as there electrolyte and applied 300 mA to a range of concentrations: 0.01, 0.03, 0.05,

0.08, 0.1, and 0.15 M. They found that higher electrolyte concentrations lead to more H2O2 with

insignificant increases over 0.08 M. They also tried a range of applied currents from 30 mA to

300mA. Once again they found that increased current resulted in increased H2O2 production up

to 150 mA after which the increase in H2O2 concentration was insignificant. Finally Soltani et al.

notes that after ten 70 minute runs there is a slight decrease in the performance of their cell,

though they do not speculate why (Soltani et al., 2012).

Assumpcao et al. (2012) did a direct comparison of two types of carbon; Printex L6 and

Vulcan XC 72R. They found that Vulcan XC 72R transferred an average of 2.9 electrons per

molecule and produced 51% H2O2. This is not particularly surprising as Vulcan XC 72R is one of

the most common carbon supports used in modern fuel cell catalysts in which noble metals are

17

loaded onto carbon supports. On the other hand, they found Printex L6 transferred an average

of 2.2 electrons per molecule and produced 88% H2O2. To determine why there was such a

difference between these two carbons they looked at the composition of each. They found sulfur,

oxygen, and nitrogen in both carbons, with the majority of it appearing as oxygenated acids.

Additionally the Printex had more than two times the oxygenated acids compared to the Vulcan.

As more oxygenated acids increases the hydrophilicity of the carbon and this facilitates the

formation of H2O2, it is believed that this is the main reason for the difference in H2O2 production

(Assumpcao et al, 2012).

Boehm et al. (1984) has also explored the effectiveness of carbon as a catalyst for ORRs.

Their group looked at 4 forms of carbon and several methods of pretreatment. The 4 forms of

carbon are: Peat charcoal, wood charcoal, sugar charcoal, and carbon black. Each catalyst was

tested in the oxidation reaction of dilute sulfuric acid by O2. It was found that peat charcoal was

a good catalyst, wood charcoal was a poor catalyst, and carbon black and sugar charcoal showed

little to no activity. Next the catalysts were subjected to heat treatment before testing. It was

found that heat treatment increased the activity of all the catalysts, with optimal temperature

treatments being 1070K for the charcoals and 1170K for the carbon black. Additionally surface

treatments where evaluated. Generally, basic surface oxidants where found to slightly increase

the activity of the catalyst and acidic surface oxidants decreased the activity of the catalyst.

However, treatment with ammonium at elevated temperatures was found to dramatically

increase the catalysts activity. It is believed that heat treatment with ammonium resulted in

Nitrogen being incorporated into the outer layers of the carbon catalyst, as seen in Fig 2.2A. The

Nitrogen in the carbon structure increases the activity of the catalyst by giving its extra electrons

18

to the conduction band which in turn gives its electrons to the adsorbed molecules, making the

carbon atoms adjacent to the nitrogen atom more active (Boehm et al., 1984).

Figure 2.2 A Models for the Substitution of Carbon by Nitrogen Atoms at the Edges of the Carbon Sheets (Boehm et al., 1984)

2.3 Proton Exchange Membrane (PEM)

2.3.1 Introduction to PEMs

Since the proposed device is based on a PEM fuel cell, a background is provided here.

Proton electron membranes (PEM) were first seen in the 1960s within fuel cells as auxiliary power

source in the Gemini space flights (Lister & McLean, 2004). Stanley H. Langer and Robert G.

Haldeman of American Cyanamid Company were able to subsequently successfully use them to

purify oxygen using 11.2 mg/cm2 of Pt on a stainless mesh screen as the electrode, while

electrolyte was simply 5 disks of filter paper saturated in 23% of KOH solution (Langer &

Haldeman, 1964). Their experimental work was the ground-work for proving that four-electron

mechanisms are operative and feasible in oxygen pumping. In addition, their work provided a

basis for other studies to be conducted such as catalytic materials and oxygen electrode

mechanisms. However, major advances in terms of redesign and configuration of PEM fuel cells

did not occur till the 1980s. Researchers have constantly been looking to enhance the design and

have succeeded by trying to reduce the expensive platinum catalyst to finding alternative

catalysts.

19

2.3.2 Mechanism

Proton exchange membranes are a type of semipermeable membranes designed to

conduct protons whiles the membranes tend to be impermeable to gases such as hydrogen or

oxygen. PEM fuel cells are based off the normal membrane electrode assembly (MEA), whose

basic function involves hydrogen oxidation at the anode, OOR at the cathode anode and transfer

of protons through the PEM.

The early electrolytes were aqueous solutions of acids and bases. To reduce the Ohmic

resistance, these could be soaked on a thin porous disk or on a membrane. The basic concept of

the polymer electrolyte involves covalently binding the acid group to a polymer in the form of a

membrane rather than immobilizing the liquid acid electrolyte within a porous support layer that

is held there physically via capillary and surface forces so that leaking of the acid is avoided. Thus,

this concept avoids the dissolution, volatility, and migration acid electrolytes altogether. To

better understand this PEM concept, it has been exemplified schematically for the case of sulfuric

acid in Figure 2.3A, where one of the –OH groups in sulfuric acid is replaced by a polymer chain

R, resulting in a solid polymer electrolyte with a sulfonic acid group, i.e., R–SO3H.

Figure 2.3 A Sulfuric Acid Analog of Polymer Electrolyte (Sulfuric Acid Model)

When such a polymer electrolyte, also called ionomeric polymer, or simply an ionomer, is brought

in contact with water, hydronium ions, or hydrated protons, are formed by the following

reaction:

20

R – SO3H + H2O ↔ H3O+ + R – SO3 (2.3.1)

The reaction facilitates conduction of protons to occur in the aqueous phase. Thus, the polymer

electrolyte acts like an ordinary acidic electrolyte, except that not only does it anchor the acid

group, the resulting anion is not solvated, and thus does not participate in conduction, which is

carried out solely by the hydronium ions. Of course, anions and hydronium ions in PEM must stay

close together to maintain overall electrical neutrality within the ionomer. These two conditions

can be satisfied only if micelles or reverse micelles are formed, with water and the polymer as

the continuous phase, respectively. Thus, reverse micelles, or inverted micelles, are formed when

water is introduced into PEMs, as for example in Nafion®.

In principle, the polymer chain R may be selected from a wide range of possible options

and, hence, a number of different PEM materials have been investigated. In practice, R must be

electrochemically and thermally stable and should preferably be hydrophobic and/or cross-linked

to avoid excessive swelling in water (Mauritz & Moore, 2004). The early polymer electrolytes

developed were blends of polymer with a highly cross-linked polystyrene-based ionomer.

However, these materials were found to not possess adequate chemical stability under the harsh

oxidative environment in an operating fuel cell, because of the instability of the C–H bond in the

polymer. A stable ionomer with excellent conductivity was found in 1962, when DuPont

developed the Nafion® membrane, which is based on a highly chemically inert backbone

structure similar to PTFE, as shown in Figure 2.3B. The chemical inertness of Nafion® is due to the

fact that the C–F bonds in it are much more stable than the C–H bonds present in the

hydrocarbon-based membranes (Coms, 2008).

21

Figure 2.3 B Chemical Structure of Nafion® (Zhou, et al 2007)

The molecular mass (weight) of Nafion® depends upon m (5-13), n (~1000), and x (0-3) (Figure

2.3B), and is typically in the range of 105 – 106 Da. The properties that are typically used to

characterize Nafion® membranes, however, are the equivalent weight (EW) and the membrane

thickness. The EW is defined as the number of grams of dry Nafion® per mol of acid groups. Thus,

it is essentially the molecular weight (Da) of the anion –RSO3−. A typical value for Nafion® is 1100.

In fact, a Nafion® membrane is denoted by a number in which the first two digits represent its

EW, while the last denotes its nominal thickness. Thus, Nafion® 117 is a membrane with an EW

of 1100 and a nominal thickness of 0.007 in (178µm). Other common membranes are Nafion®115

(0.005 in. or 125µm) and Nafion® 112 (0.002 in. or 50µm) (Sigma Aldrich). While a thinner

membrane can provide better fuel cell performance due to lower resistance, it is less durable

than a thicker membrane and has a higher permeability to O2 and H2, resulting in more crossover.

The PFSA backbone is strongly hydrophobic, while the proton conducting sulfonic acid group is

highly hydrophilic and, thus, phase separation readily occurs when water is introduced into the

Nafion®, forming interconnected aqueous reverse micelles or clusters, roughly 4 nm in size and

interconnected by channels of roughly 1 nm size responsible for percolation, that contain water,

the anion, and the hydronium ions, as shown in Figure 2.3C, (Personal Notes by Prof. Datta).

22

Figure 2.3 C Reverse micellar Cluster-Network Structure of Hydrated Nafion®

By minimizing the interfacial area, the spherical shape for clusters or inverted micelles ensures

minimum energy of interaction between the hydrophobic and the hydrophilic regions.

Another interesting aspect of these membranes is the anomalously high conductivity of

hydronium ions. For instance, at 25 ºC, λ1 0 = 349.8 S.cm2/equiv in water, which is extremely high

when compared with equivalent conductivity of other cations of size similar to the hydronium

ion, e.g., Na+. In other words, the conductivity cannot be accounted for simply by the

hydrodynamic vehicular diffusion, in which the hydronium ion diffuses en masse, as modeled, for

instance by the Stokes-Einstein model. The difference can be explained by an unusual mechanism

known as the Grotthuss (or structural) mechanism of proton diffusion that supplements the en

masse diffusion. It was proposed over two-hundred years ago, prior to the formulation of Fick’s

law, and imagines that the proton simply hops from the hydronium ion to an adjacent water

molecule, becoming a hydronium ion and leaving a water molecule behind, and so on, as shown

schematically in Figure 2.3D. The Grotthuss mechanism involves two sequential steps, namely,

rotation of the dipolar water molecule due to the electric field of the adjacent hydronium ion

into a receptive orientation, followed by the transfer of proton to the water molecule, via

quantum mechanical tunneling from hydronium ion. The Grotthuss model of “chain mechanism”

23

for the transfer of protons in water is ingenious considering that in 1806 the chemical formula of

water was not settled, and the concept of molecules was new (Personal Notes by Prof. Datta)..

Figure 2.3 D Proton Diffusion via (a) en masse or vehicle and (b) Grotthuss Mechanism.

2.3.3 Hydrogen Oxygen Fuel Cell

The design and structure of fuel cell is described below. The electrode is the layer that sits

on each side of the PEM. Therefore, to ensure an effective design, an electrode must provide for

the three main transport processes with the fuel cell. These transport processes include protons

from the membrane to the catalyst, electrons from current collectors to the catalyst through the

gas diffusion layer, and finally the reactant and product gases to and from the catalyst layer and

the gas channels. Protons, electrons and gases are known as the three phases that can be found

within the electrocatalyst layer. These phases must be correctly spread across the catalyst layer

to optimize that electrode design and reduce transport loss. Thus, an effective line plate interface

among between phases is needed for the PEM fuel cell to operate effectively.

24

The PEM is the central piece of the membrane electrode assembly (MEA) and exterior

layers on each side together form the electrode. The next layer is the Catalyst Layer as seen in

Figure 2.3E sits between membrane and gas diffusion layer (GDL), and is also known as the active

layer (Lister & McLean, 2004). The layer is the location where the half-cell electrode reaction takes

place in the PEM fuel cell. Adjacent to the catalyst layer, as seen in Figure 2.3E, is the gas diffusion

layer whose sole purpose is to ensure reactants diffuse effectively to the catalyst layer and also

transports electrons to and from the catalyst layer. In most cases, the layer is made up of porous

carbon paper or a graphite cloth, which is roughly 100-300 um thick (Lister & McLean, 2004).

Figure 2.3 E Sample Proton Exchange Membrane Fuel Cell Schematic (Lister & McLean, 2004)

25

PEMFC is fueled by hydrogen and the charge carrier is the hydrogen ion (H+). At the anode, the

hydrogen molecule is split into hydrogen ions and electrons. These hydrogen ions permeate

across the electrolyte to the cathode side. However, the electrons flow through an external path

and generate electrical power. Oxygen, usually in the form of air, is supplied to the cathode and

combines with the electrons and the hydrogen ions coming above the electrolyte layer to

produce water. The reactions at the electrodes are as follows:

Anode reaction: 2H2 → 4H+ + 4e-

Cathode reaction: O2 + 4H+ + 4e- → 2H2O

Overall cell reaction: 2H2 + O2 → 2H2O

2.3.4 Oxygen generation case studies

The following section will address many advances in this field of research. Significant

advances were not conducted until the 1980s. For instance, Yuko Fujita et al. (1985) conducted

a research study on an oxygen separator based on oxygen reduction at the air cathode. This study

specifically focused on the previous O2 separators that used a liquid electrolyte, and that there

was little published work on O2 separators that use a polymer electrolyte membrane. Thus, was

one of the first published studies using the Nafion® 117. Pt anode was plated onto the membrane

and Pt/C cathode was hot pressed onto the membrane. In addition, the series of experimental

testing was conducted on 10 cm2 and 100 cm2 active areas. Fujita et al. (1985) found the following

parameters which included operating temperature of 40o C water flowing to the anode and air

at STP that flows 4 L/min to the cathode (Fujita, 1985). With this setup, the output of the PEMFC

is 70.9 ml/min O2, with a concentration of 98.4%, and current efficiency of oxygen reduction or

φ of 91.3%.

26

The test method was to run on circulation mode in order to remove all O2 from 1 L of air.

It took 70 min to remove all O2 (final O2 concentration of 0.02%) as seen in Figure 2.3F. At this

scale, in flow through mode, a stream of 0.02% O2 can be achieved with a flow of 100 ml/min.

Figure 2.3 F Graphical representation for the oxygen concentration over time (Fujita, Nakamura & Muto, 1985)

Based on the Fujita study, the following conclusions can be made that: (a) water produced

at cathode is discarded, (b) water at anode is recycled and (c) air feed is humidified at 40o C. For

longevity tests, Figure 2.3G shows the 100 cm2 cell was used intermittently for 7.5 hours per day

at 200 mA/ cm2. In addition, the overall design is an effective O2 separator and includes the

following: lower φ and YO2 than Takenaka et al (1982), superior O2 separation than liquid

electrolyte systems, no decrease in performance over 100 days as seen in Figure 2.3G and that

humidified air doesn’t change the cells operation. Additional observations are that humidified air

27

doesn’t flood the cathode and recycling the water from the cathode doesn’t seem to build up

impurities.

Figure 2.3 G Change in voltage of the PEM Fuel Cell over time (Fujita, Nakamura & Muto, 1985).

Furthermore a more recent study conducted for O2 separation with PEM technology is

described next. The removal of O2 from the air is important as low O2 levels help prevent food

spoiling, metal corrosion, and are needed for some biological reactors. There are several non-

electrochemical ways to lower the O2 levels: adding N2, selective O2 combustion, O2 selective

reduction, selective adsorption, or membrane separation. However, there are more effective

electrochemical ways to remove O2 from air. Winnick (1990) reviewed electrochemical O2

extraction and Langer & Haldeman (1964) was able to recover pure O2 from air in alkaline and

acidic solutions. The equilibrium voltage of such a cell is 0, but the actual cell voltage is equal to

the 2 over potentials, plus the Ohmic drop across the cell. Additionally, General Electrics has

recently developed a similar system to provide O2 from high pressure air. Several patents also

28

describe processes where O2 is reduced to O2- and transported through an oxide lattice that is

missing oxygens (i.e. a solid oxide conductor). However, the current densities are below 100

mA/cm2 and the operating temperature is above 500o C. Tseung & Jasem (1981) imagined O2

extraction by its reduction to hydrogen peroxide on a graphite based cathode through 2 electron

ORR. Brillas et al. (1997), developed a 2 electron reduction path across a membrane using a

NiCo2O4 surface to reform O2 (current density is limited below 0.2 A/cm2 by the finite

concentration of peroxide). Recently using proton exchange membrane fuel cells (PEMFC), O2

extraction was reached with current densities of 0.6 A/cm2 and 0.015 M O2 per second per m2

membrane. Based on this information, Eladeb et al. (2012) conducted a series of experiments on

PEM fuel cells to garner a better understand on the performance of a PEMFC. Their goal was to

remove all O2 from a stream of air using PEM technology. To achieve this they used a standard

water electrolysis MEA, as seen in Figure 2.3H, to reduce oxygen from the inlet stream at the

cathode to water. The formed water is then electrolyzed at the anode yielding oxygen, protons,

and electrons. The protons are then recycled within the membrane for subsequent oxygen

reduction at the cathode (Eladeb et al. 2012).

The MEA was designed for water electrolysis and experiments took place between 50o

and 80o C. Liquid water was pumped to the anode compartment and heated before entering the

cell. The cell was operated at a fixed voltage or controlled current density using a PGSTAT

30autolab potensiostat connected to a 20 amp autolab booster (Eladeb et al., 2012).

29

Figure 2.3 H Schematic view of the 25 cm2 cell for oxygen extraction (Eladeb et al., 2012)

The current density vs. voltage was established in either potensiostatic or galvinostatic

modes so that the cell potential was below 1.4 v for long runs. Most measurements were carried

out with air but O2 was also used for comparison. O2 extraction was achieved repeatedly with a

fixed current density and cell voltage less than 1.4 v. During most runs in this study the cell voltage

stabilized after 10 to 30 minutes. Outlet gas composition was determined with gas

chromatography. The following results are for all experimental runs at T = 60o. Figure 2.3I shows

that the air flow rate affects the current density at 1.3 v. with the current density increasing with

the increase in flow rate (Eladeb et al., 2012).

30

Figure 2.3 I Current Density vs. Air Flow (Eladeb et al., 2012).

Figure 2.3J shows that the cell current density is an increasing function of the

voltage. These results are consistent with other research. Voltage and current density are

proportional. And we see better efficiency with pure O2. Figure 2.3K shows the stability of the

cell for long runs of up to 50 hours (Eladeb et al., 2012).

Figure 2.3L demonstrates that as inlet O2 flow increases the outlet O2 flow approaches

atmospheric composition (21% O2). This is expected and demonstrates the optimal flow rate for

the cell (lambda ~ 5-10). Furthermore, this graph indicates that at larger flow rates calculating

efficiency of the ORR will be imprecise, as demonstrated in Figure 2.3L. Despite low precision,

Figure 2.3L clearly shows a decrease of ORR efficiency as lambda increases. Furthermore, the

graph shows that the average ORR efficiency is between 70% and 100% (Eladeb et al., 2012).

31

Figure 2.3 J Current Density vs. Cell Voltage (Eladeb et al., 2012)

Figure 2.3 K Current Density vs. Time (Eladeb et al., 2012)

32

Figure 2.3 L O2 Outlet Flow vs. the Inlet O2 Flow over Current (Eladeb et al., 2012)

Figure 2.3M demonstrates that for current densities of 100 mA/cm2 or more a 90% or

higher ORR efficiency can be observed. The graph shows that current density has a positive effect

on ORR efficiency. Thus, this study proves the validity of using water electrolysis in a PEM fuel

cell like device for oxygen pumping (Eladeb et al., 2012).

33

Figure 2.3 M Current Density vs. the ORR efficiency (Eladeb et al., 2012)

2.4 Anion-Exchange Membrane (AEM)

2.4.1 Development

Alkaline fuel cells were developed in the 1930s by F. T. Bacon (Arges et al., 2010). Initial

fuel cells used a liquid electrolyte, commonly an aqueous solution of KOH due to its effectiveness

as a highly conducting alkaline hydroxide (Merle et al., 2011). These fuel cells suffered from

problems due to the use of the liquid electrolyte. The strong alkaline electrolytes result in its

reaction with carbon dioxide in air, resulting in the formation of carbonates and the

34

corresponding reduction of the connectivity of the electrolyte (Arges et al., 2010). This

deterioration is discussed further in the subsequent sections.

The recent development of anion exchange membranes has eliminated the need for a

liquid electrolyte and promoted the use of an anion exchange membrane as the medium for the

transport of hydroxide ions. In these membranes, as in proton exchange membranes, the

electrolytes are fixed to polymer chains (Arges et al., 2010). Progress on anion exchange

membranes and their use in fuel cells lags behind that of the proton exchange membranes.

Commercial production of anion exchange membranes has only begun recently, and research

concerning these membranes is currently being conducted. These membranes have proven to

resist contamination and maintain ionic conductivity in neutral environments over an extended

period of time (Vega et al., 2010). Such studies show that the anion exchange membranes are

able to operate at a reasonable level, even in ambient air, and are able to be a viable option to

the proton exchange membrane.

2.4.2 Mechanism

An anion exchange membrane consists of a fixed polymer backbone on which electrolytes

that have interchangeable anions are attached. Common anion exchange groups in anion

exchange membranes are quaternary ammonium, quaternary phosphonium, and tertiary

sulfonium (Merle et al., 2011) In the presence of a solvent, these fixed polymers become mobile

and are able to transfer charge. Like in a proton exchange membrane, ion transport in an anion

exchange membrane occurs via the Grotthuss mechanism. In this mechanism, hydroxide ions are

transported through the membrane along a chain of water molecules. The ion moves from one

35

molecule to another by means of the formation and cleavage of hydrogen bonds as seen in Figure

2.4A, making a tetrahedral water molecule with each bond formed (Merle et al., 2011).

Figure 2.4 A Schematic of dissociation and solvation of the pendant OH- groups within the pores of a hydrated AEM (Grew et al., 2010)

Anion exchange membrane fuel cells function in a very similar way to proton exchange

membrane fuel cells. The difference between the two lies in the ions that are transferred

between the cathode and the anode. Within the anion exchange membrane fuel cell, hydroxide

ions are produced through oxygen reduction at the cathode. This hydroxide ion is carried through

the membrane by way of the polymer electrolyte. Upon reaching the anode, the hydroxide ion

combines with hydrogen and form water. The electrons that are displaced during the hydrogen

oxidation return to the cathode where they participate in the reduction of oxygen that produces

the hydroxide ions. The flow of electrons from the anode to the cathode provides the electrical

energy that is produced by the fuel cell.

36

2.4.3 Benefits and Challenges

Anion exchange membranes have numerous advantageous aspects to their use. Among

these beneficial qualities are the capabilities of these membranes in fuel cells to operate using a

variety of fuels, generate high energy density, and they can be operated at moderate

temperatures. One of the largest benefits to using anion exchange membranes is the expensive

metal catalysts used in proton exchange membrane fuel cells are not essential for operation. This

is due to the higher reaction kinetics at the electrodes in the alkaline conditions of the anion

exchange membrane, especially for the oxygen reduction reaction, which translates into a higher

electrical efficiency and a lower cost (Merle et al., 2011). This allows either the use of a less

expensive catalyst or a lesser amount of the traditional platinum catalyst.

In general, many of the anion exchange membranes benefits are related to the reducing

the overall costs of operating a device using an anion exchange membrane. The ability to operate

at a lower temperature means less energy is required to maintain the unit at the desired

temperature. The ability to use a less expensive catalyst or a smaller amount of an expensive

catalyst also lowers the required costs associated with the process. While these benefits have

some promising attributes, several problems with the use of anion exchange membranes exist.

One issue with anion exchange membranes include the inability for hydroxide ions to

dissociate as strongly as hydrogen ions. While this can be enhanced with the increasing of the

number of cationic sites, this method ultimately leads to the deterioration of the membrane itself

(Arges et al., 2010). Another problem with the anion exchange membrane is the susceptibility of

the hydroxide groups to be neutralized by carbon dioxide. When the membrane is exposed to

air, the hydroxide ions are replaced with carbonate and bicarbonate ions (Varcoe et al., 2010),

which in turn reduces the pH. The reduction in pH slows the kinetics of the reactions at both the

37

cathode and the anode, and the larger carbonate and bicarbonate anions cause a decrease in the

conductivity of the membrane. This causes a decline in the efficiency and performance of the

membrane (Arges et al., 2010).

Another challenge facing anion exchange membranes at this time is the lack of research

and development in the area. Of course, this problem will solve itself over time as more research

concerning these membranes is conducted. At this time however, anion exchange membranes

fall far behind proton exchange membranes in terms of use in membrane electrode assemblies

(Arges et al., 2010). In addition, much of the research concerning anion exchange membranes

has occurred at the traditional conditions, including the use of expensive catalysts such as

platinum. In essence, many of these studies have not taken advantage of the inherent benefits

of an anion exchange membrane. Further studies that examine the anion exchange membrane

in more appropriate conditions are required to accurately compare the capabilities of the

membrane compared to that of the proton exchange membrane in fuel cells.

2.4.4 Oxygen Generation Case Studies

Due to the relatively young age of the anion exchange membrane, there have not been

any studies exploring the generation of oxygen using an anion exchange membrane. Considering

the research conducted on the generation of oxygen using a proton exchange membrane and the

recent developments in developing stable anion exchange membranes, it is very likely that

studies exploring this in anion exchange membranes will soon come to light. With the numerous

advantages of anion exchange membranes, including the use of cheaper catalysts and the ability

to function at lower operating temperatures, anion exchange membranes appear to be a possibly

very effective component in oxygen generation.

38

3. Methodology

3.1 Conceptual Design

As previously discussed in the literature review, there are several applications in the

medical field for an oxygen concentrator or an oxygen pump. Current solutions to this problem

are both large and inconvenient or contain compressed gas which could potentially be

dangerous. For this reason, it was decided to pursue the development an oxygen pump that

operates electrochemically using an ion exchange membrane in a fuel cell like device. Given the

nascent technology for anion exchange membranes, it was decided to pursue models based on

proton exchange membranes.

Figure 3.1 A Electrolysis Aided PEM Pump Model

Two potential oxygen pump transport models were conceived. The first model is an

electrolysis aided oxygen pump. A schematic of this system can be seen in Figure 3.1A. In this

model, oxygen is reduced from the inlet stream at the cathode to water. The formed water is

then electrolyzed at the anode yielding oxygen, protons, and electrons. The protons are recycled

39

within the membrane for subsequent oxygen reduction at the cathode. The oxygen formed at

the anode is released in the anode exhaust stream as a part of the oxygen enriched stream.

Figure 3.1 B Two Electron Oxygen Reduction Reaction PEM Pump

The second conceptual model for oxygen transport is by way of the two electron oxygen

reduction reaction. A schematic of this process can be seen in Figure 3.1B. Oxygen from the

cathode inlet is reduced to hydrogen peroxide at the cathode. Hydrogen peroxide is transported

through the membrane and is electrolyzed at the anode, yielding oxygen, protons, and electrons.

The protons are recycled within the membrane, and the oxygen is released in the anode exhaust

stream as part of the oxygen enriched stream.

40

3.2 Apparatus

Figure 3.2 A Example Schematic of Experimental Procedure

Figure 3.2 B Diagram for the fuel cell where the cathode is receiving the humidified inlet while the anode is receiving a dry inlet. In addition, the fuel cell is connected to the power supply and the Fuel Cell Test System

41

The MEA was housed in the fuel cell assembly designed for a 5cm2 MEA as seen in the

Figure 3.2B and a sample process flow diagram has been provided in Figure 3.2A. Temperature

controlled humidifiers accompanied the test station to provide desired humidified inlets to the

fuel cell. In addition, both the anode and cathode were connected to the waste stream to ensure

no release of gases to the environment. Swagelok connections were used to ensure no leaks

occurred during the experimental runs. The testing station used to conduct the experimental

runs was located in the Fuel Cell Center at Worcester Polytechnic Institute as seen in Figure 3.2C.

Figure 3.2 C Overall Diagram of the Fuel Cell Test bed located in the Fuel Cell Center at Worcester Polytechnic Institute

Additionally, the test station contained the Fuel Cell Test System Series 89B, alongside a

computer interface which plotted the data recorded. This system, was also used to apply small

amounts of current (mA) during the activation of the each MEA. The Handi+, a portable battery-

powered oxygen sensor manufactured by Maxtec, was used to measure the O2 concentration in

42

the product stream with an accuracy of 0.1%. The sensor was calibrated before each MEA that

was tested using the oxygen or air tank in Goddard Hall’s Fuel Cell Center. The test bed shown in

Figure 3.2C contains other pieces of equipment such as valves, piping and pressure gauges which

were not necessarily a part of these experimental runs.

The HP 6651A power supply, as seen in Figure 3.2D, was used to apply voltage across the fuel cell

during the experimental runs.

Figure 3.2 D HP power supply used for the experimental runs

3.3 Materials

The experimental procedures in this study were carried out using various membrane

electrode assemblies (MEAs). Each MEA consisted of a Nafion® 115 membrane, a catalyst layer

on each side of the membrane, and a gas diffusion layer on each side of the membrane. All of the

MEAs tested had an active transport area of 5 cm2. Two types of MEAs were used for oxygen

transport studies, one with Printex L6 cathode loading and 1 mg Pt/C, the other with 1 mg Pt/C

loading on both the cathode and the anode. An MEA for electrolysis was also used in studies,

with 1 mg Pt/C loaded at the cathode and 3 mg PtIrB loaded at the anode. A complete summary

of the MEAs testing in this study can be seen in Table 3.3A.

43

Table 3.3 A Summary of Tested MEAs

MEA Name Membrane Cathode Catalyst

Anode Catalyst

Product # Supplier

P1 Nafion® 212 0.5 mg/cm2 Pt

on carbon 0.5 mg/cm2 Pt

on carbon 5L HP-A MEA

5cm2 FuelCellStore

P2 Nafion® 212 0.5 mg/cm2 Pt

on carbon 0.5 mg/cm2 Pt

on carbon 5L HP-A MEA

5cm2 FuelCellStore

P3 Nafion® 115 0.1 mg/cm2 Pt

on carbon 0.1 mg/cm2 Pt

on carbon Custom 5cm2

MEA FuelCellsEtc

C1 Nafion® 115 1 mg/cm2 Printex L6

0.1 mg/cm2 Pt on carbon

Custom 5cm2 MEA

FuelCellsEtc

C2 Nafion® 115 1 mg/cm2 Printex L6

0.1 mg/cm2 Pt on carbon

Custom 5cm2 MEA

FuelCellsEtc

I1 Nafion® 115 0.1 mg/cm2 Pt

on carbon 3 mg/cm2

PtIrB Custom 5cm2

MEA FuelCellsEtc

I2 Nafion® 115 0.1 mg/cm2 Pt

on carbon 3 mg/cm2

PtIrB Custom 5cm2

MEA FuelCellsEtc

In order to study the transport of hydrogen peroxide across the membrane, a 35%

solution of hydrogen peroxide was also obtained from Alfa Aesar. This was used to saturate inlet

streams with hydrogen peroxide in an effort to facilitate hydrogen peroxide transport through

the membrane.

Various additional materials were used in these experimental procedures. Rotameters

were purchased from Cole Parmer and had a flow range from 0 to 3 L/min. Compressed gas tanks

containing air, oxygen, hydrogen, and helium were used as the feed to the system. All tubing and

fittings were purchased through Swagelok.

3.4 Experimental

The first series of MEAs tested were the standard platinum supported on carbon. To

activate these, humidified hydrogen was fed to the anode and humidified air to the cathode for

a period of 1 – 3 hours. Following this, a current of 0.5 – 0.6 A was applied for 4 – 5 hours. The

final step in the activation was to apply varying currents between 0.4 A and 0.8 A for 15 – 20

44

minutes each, a total of four times. For the first MEA an ATR-IR spectrum was also taken in-

between each step of the activation. At the end of activation a polarization curve was taken by

scanning the current from 0 to 3 A in 0.05 A increments, holding each current for 60 seconds.

For MEA P1, the first set of experiments involved feeding humidified air to both sides of

the cell and applying a potential of 1.4 V for 20 minutes, then 1.5 V for 30 minutes. The next set

of experiments was to feed humidified He to both sides of the cell and apply 1.5 V for an hour,

then 1.6 V for 45 minutes. The final experiment was to feed heated humidified He to both sides

of the cell, this was accomplished by heating the humidifiers from 11o C to 27o C before beginning

the experimental run. Once the humidifiers were at 27o C a potential of 1.6 V was applied for 5

minutes, immediately followed by a potential of 1.8 V for 5 minutes, immediately followed by a

potential of 2.0 V for 1 hour. After the first 40 minutes of the experiment the humidifiers were

heated to 50o C. At the conclusion of these tests a polarization curve was taken, as described

above at the end of the activation, and compared to the first polarization curve to determine the

condition of the MEA.

For MEA P2, the first set of experiments involved feeding humidified He to the cathode

and dry He to the anode while applying potentials between 1.4 V and 2.0 V for 5 minute intervals.

The next set of experiments was to feed humidified air to the cathode and dry air to the anode

while applying various potentials between 1.3 V and 2.0 V for 5 minutes apiece. The final set of

experiments was to feed dry air to the anode and dry O2 to the cathode while applying potentials

between 0.8 V and 2.0 V for periods of time shorter than 10 minutes. At the conclusion of these

45

tests a polarization curve was taken, as described above at the end of the activation, and

compared to the first polarization curve to determine the condition of the MEA.

For MEA P3, only two experiments were run. First He with H2O and H2O2 vapor was fed

to the cathode and humidified air was fed to the anode while applying potentials between 1.0 V

and 1.2 V. Then humidified air was fed to the cathode and He with H2O and H2O2 vapor was fed

to the anode while applying potentials between 1.0 V and 1.2 V.

The next series of MEAs tested were the Printex L6 (carbon). To activate these, humidified

air was fed to both sides overnight, 12 – 20 hours. For MEA C1, the first set of experiments

involved applying potentials between 0.4 and 1.0 V while first feeding humidified air to both sides

and then dry air to both sides. The next set of experiments was to feed H2 to the Pt anode and

humidified air to the carbon cathode while varying the applied potential. The last set of

experiments preformed fed humidified O2 to the cathode and dry air to the anode while varying

the applied potential between 1.0 V and 0.4 V.

For MEA C2, the first set of experiments involved feeding humidified O2 to the cathode

and dry air to the anode while applying varying potentials between 1.3 V and 0.4 V. The next set

of experiments involved feeding first dry air then humidified air to the anode and He with H2O

and H2O2 vapor to the cathode while varying the applied potential. The third set of experiments

was to feed He with H2O and H2O2 vapor to the anode and humidified air to the cathode while

varying the potential.

The last series of MEAs tested were standard electrolysis MEAs that had solid Pt/Ir anodes

and Pt supported on carbon at the cathode. To activate these MEAs, humidified air was fed to

46

both sides for an hour. Next humidified air was fed to both sides for 4 – 5 hours and a constant

potential of 0.4 V to 0.5 V is applied. The last step was to apply varying potentials between 0.2 V

and 0.8 V for 20 minute periods while feeding humidified air to both sides. For MEA I1, the first

set of experiments was to feed humidified air to both sides of the cell while varying the applied

potential between 1.2 V and 1.8 V and varying the systems temperature between 11o C and 80o

C. The next set of experiments was to apply potentials between 1.0 V and 1.2 V while first feeding

humidified air to the cathode and He with H2O and H2O2 vapor to the anode; then switching the

feeds so that the He with H2O and H2O2 vapor is going to the cathode and the humidified air is

going to the anode. The final set of experiments involved feeding humidified O2 to the cathode

and liquid water to the anode while applying 1.4 V to 1.5 V.

The last membrane tested, MEA I2, was tested in four general areas. The first was testing

H2O2 electrolysis by feeding He with H2O2 vapor and H2O vapor to the anode and feeding

humidified air to the cathode while applying potentials between 0.8 V and 1.5 V. The second test

was performing liquid water electrolysis rather than vapor electrolysis, and was tested by feeding

liquid water to the anode and feeding humidified He to the cathode while applying various

potentials between 1.6 V and 2.5 V. The third test was vapor electrolysis, which was tested by

feeding both humidified He and air to both sides in turn while applying potentials between 1.2 V

and 2.0 V. The final test was an electrolysis driven oxygen pump which was tested by feeding

each in turn: dry O2, humidified O2, humidified air, and humidified He to the cathode and feeding

each in turn: humidified He, humidified air, and liquid water to the anode while applying

potentials between 1.2 V and 2.0 V. For the full tabularized list of tests and results on all MEAs,

please refer to Appendix A.

47

4. Results & Discussion

4.1 Liquid Electrolysis

This set of experiments attempted to examine water electrolysis on a proton exchange

membrane as a method of transporting oxygen across the membrane as shown in Figure 3.1A.

All of the water electrolysis runs were performed with an MEA loaded with Pt/C at the cathode

and unsupported PtIrB at the anode. As discussed previously, this process involves the oxidation

of water to produce oxygen and hydrogen. Initially, tests were performed in an attempt to

perform electrolysis in the MEA. As seen in Table 4.1A, the largest measurable current during

these trials was 108 mA. These results were gathered feeding liquid water to the anode of the

cell at 10 mL/min. While the oxygen produced was not measured, the presence of gas in the

liquid outlet stream and a current signifies that electrolysis is occurring at the membrane.

Electrolysis is observed at voltages higher than 1.23 V, which is the theoretical voltage for water

electrolysis. As seen in the results, the current density of the cell increases as voltage increases,

and a maximum current density is not observed.

Table 4.1 A MEA I2 Liquid Electrolysis Results

Applied Voltage Current Humidifier

Temperature Cell Temperature

1.6 V 40 mA 60 C 60 C

1.8 V 67 mA 60 C 60 C

2.0 V 81 mA 60 C 60 C

2.2 V 94 mA 60 C 60 C

2.5 V 108 mA 60 C 60 C

48

The measured current density for water electrolysis in this study is, however, inexplicably

much lower than published values for proton exchange membrane electrolysis (Eladeb et al.,

2012; Greenway et al., 2009). A potential cause for this problem is the formation of vapor bubbles

on the membrane at the site of water electrolysis. The formation of vapor bubbles in the liquid

feed to the cell can seriously hinder the mass transfer and hence the performance of the

electrolysis cell. While flowing bubbles can potentially provide transport of liquid water to the

membrane, too many bubbles can limit the contact area between the liquid water and the

surface of the membrane (Spurgeon & Lewis, 2011). This in turn reduces the amount of

electrolysis that occurs. It’s very likely that a limiting density can be reached, with the amount of

bubbles forming on the membrane limiting any increase in electrolysis, although such a limit was

not reached. The flow rate of liquid water may also have an effect on the formation of bubble at

the site of electrolysis. Due to the lack of a settable pump, this effect was unable to be measured.

Many studies opt for low liquid water flow rates when performing proton exchange membrane

electrolysis, although a higher flow rate may be more beneficial due to the likelihood of the flow

either pushing the forming vapor to the membrane or flushing it out of the system. It has been

shown that a higher stoichiometric ratio of water, associated with a higher flow rate, decreases

the current density in an electrolysis membrane (Greenway et al., 2009). It follows that proper

operation requires the correct balance between these two parameters. A more in-depth and

precise testing procedure would most likely provide a more favorable result.

4.2 Vapor Electrolysis

Vapor electrolysis was also examined in this study using a similar membrane, a Pt/C

catalyst at the cathode and unsupported PtIrB catalyst at the anode. The mechanism is exactly

49

the same as liquid electrolysis, with the exception that water vapor is used as the feed to the

system rather than liquid. More trials were attempted to create an oxygen pump using the same

membrane. The process would include oxygen reduction at the cathode and water electrolysis at

the anode to pass oxygen from the cathode to the anode as seen in Figure 3.1A. In these trials,

the current density and the change in oxygen composition were measured to assess the

performance of the cell. As seen in the results, these trials provided less than desirable results.

At most, the oxygen content of the outlet stream rose 0.3%. A summary table of the complete

results of these runs can be found in Appendix A.

Table 4.2 A Vapor Electrolysis Results at Ambient Temperature

Cathode Feed Anode Feed Applied

Voltage

Observed

Current

Change in Oxygen

Concentration

Humidified Air at

60 mL/min

Humidified Air at

40 mL/min

1.3 V 26 mA 0.05 %

1.4 V 40 mA 0.1 %

1.5 V 47 mA 0.1 %

1.6 V 53 mA 0.1 %

1.7 V 60 mA 0.1 %

1.8 V 67 mA 0.1 %

50

Table 4.2 B Vapor Electrolysis Results at Elevated Temperature

Cathode Feed Anode Feed Applied Voltage

Observed Current

Change in Oxygen Concentration

Humidifier Temperature

Humidified Air

at 235 mL/min

Humidified Air

at 235 mL/min

1.5 V 94 mA 0.2 % 61 C

1.5 V 94 mA 0.3 % 65 C

1.5 V 108 mA 0.1 % 71 C

1.5 V 163 mA 0.2 % 75 C

1.5 V 163 mA 0.2 % 80 C

As seen in the results, vapor electrolysis yields similar current densities to that of liquid

electrolysis. While vapor electrolysis avoid the problem of bubble formation which was discussed

earlier, there are a number of issues with vapor electrolysis. The majority of these problems stem

from the mass transport limitations that occur at higher current densities with water vapor

(Spurgeon & Lewis, 2011; Greenway et al., 2009; Fox & Colón-Mercado, 2011). A mass flux limit

is reached at relatively low values of electrolysis, as water molecules are unable diffuse any faster

through the membrane.

This mass flux limit could potentially be caused by two things. The first possible issue with

the system is the formation of water in the MEA. It is known that liquid water can be detrimental

to the operation of a PEM fuel cell, as the excess presence of water can smother the gas

electrodes and ultimately flood it (Pasaogullari & Wang, 2004). This can also be a serious issue at

the gas diffusion layer as well (Pasaogullari & Wang, 2004; Litster et al., 2006). When considering

51

these studies along with the conceptual model of the electrolysis aided oxygen pump, this could

be an issue for the oxygen reduction reaction. There is the possibility that too much water could

completely flood the gas diffusion layer. If too much water is created at the cathode and the gas

diffusion layer is also flooded, oxygen from the cathode inlet stream is prevented from passing

through to the catalyst. This could potentially be a cause for a mass flow limitation in the system.

While the presence of liquid water in the membrane is beneficial to the system, the level at which

it would have to be controlled may be an issue that prevents the design of an efficient system

based on an electrolysis aided pump.

Oxygen transport concerns through the gas diffusion layer is a topic of concern as well. It

has been shown that the presence of nitrogen in the cathode feed stream significantly reduces

the transport of oxygen across the gas diffusion layer (Benzinger et al., 2011). The presence of

the nitrogen takes up space in the gas diffusion layer and inhibits oxygen transport. This issue is

not unique to this design, as it is also an issue for air-fed PEM fuel cells. Normally operating fuel

cells do not seem to have a serious issue with this problem, so it is possible that this should be of

no concern in this case. It is more likely that this oxygen transport issue, when combined with

the liquid water problem discussed earlier, is a possible hindrance to the operation of the

electrolysis aided pump.

Relative humidity is another issue with vapor electrolysis, as the feed streams must be at

a relative humidity of 95% or higher, otherwise the electrolysis activity is greatly diminished

(Spurgeon & Lewis, 2011). It is assumed here that the relative humidity of the streams in this

experiment were of adequate values, although the humidity was never measured. This signifies

52

that the mass flux limitation is the limiting factor to these trials. Published results show that the

voltage at which this limit is reached lies between 1.6 and 2.0 V, at which the current density at

room temperature is varying between 40 mA/cm2 and 90 mA/cm2 (Spurgeon & Lewis, 2011;

Greenway et al., 2009). Using these numbers, the amount of oxygen that could be produced

through vapor electrolysis is roughly 1.6 mL/min. The efforts to scale this up to a practical design

would not be worthwhile, as the system would require too large of a housing, making it

impractical for the intended use as put forth by this study. Although vapor electrolysis can be

used to produce oxygen, its mass flow limitations at higher current densities and its requirement

for high relative humidity prove it to be impractical.

4.3 Two Electron Oxygen Reduction

This group of experiments examined the two electron oxygen reduction aided O2 pump

(as seen schematically in Figure 3.1B). The first set of experiments were on the carbon (Printex

L6) catalyzed MEAs. MEA C1 was tested by applying varying potentials between 0.4 V and 1.0 V

while feeding different gases to the cell. We tested dry air to both sides, humidified air to both

sides, and humidified O2 to the cathode with dry air to the anode. No current or change in percent

O2 where observed for all trials. This lack of generated current indicates that the desired 2 e- ORR

and O2 transfer was not achieved, as further supported by the static O2 levels observed during

the tests. After these tests, MEA C1 was run in fuel cell mode by feeding humidified O2 to the

cathode and dry H2 to the anode resulting in an OCV of 0.20 V. This OCV indicates a very low level

of electrochemical active and as such, prompted the end of testing on MEA C1.

MEA C2 was tested by applying potentials between 0.4 V to 1.3 V while feeding humidified

O2 to the cathode with dry air to the anode. No current or change in O2 levels were observed

53

below 1.2 V. However, small currents of 12 mA and 26 mA were observed at 1.2 V and 1.3 V

respectively, with no discernable change in percent O2 in the anode exhaust. The complete lack

of activity below 1.2 V indicates the desired O2 transfer was not achieved. Additionally the small

currents achieved at and above 1.2 V show that MEA C2 is electrochemically active and capable

of H2O2 electrolysis. Next MEA C2 was tested by applying potentials between 1.0 V and 2.5 V

while feeding humidified air to the anode with He bubbled through 35% H2O2 fed to the cathode.

Below 1.2 V there was no observable current or change in O2 levels. At higher potentials an

increase in current and decrease in O2 at the anode exhaust was observed up to 135 mA and –

4.2% O2 at 2.5 V, once again indicating that MEA C2 is electrochemically active and capable of

H2O2 electrolysis. The final testing for MEA C2 involved applying potentials between 1.0 V and

1.2 V while feeding humidified air to the cathode with He bubbled through 35% H2O2 to the

anode. For these final trials no change in current or O2 levels where observed.

The next MEA tested was a standard fuel cell MEA with Pt supported on carbon as the

catalyst for both sides. First varying potentials between 1.0 V and 1.2 V were applied while

humidified air was fed to the cathode and He bubbled through H2O2 was fed to the anode. No

change in O2 levels or current was observed. Next He bubbled through H2O2 was fed to the

cathode and humidified air was fed to the anode while potentials between 1.0 V and 1.2 V were

applied. Each of these trials generated a small current and a small decrease in O2 levels, as seen

in Table 4.3A. This small decrease in O2 levels at the anode coupled with potentials under 1.2 V

suggest that we are electrolyzing the supplied H2O2 at the cathode and transporting H+ across the

membrane to form H2 at the anode.

54

Table 4.3 A Test result for MEA P3 with He bubbled through H2O2 fed to the cathode and humidified air fed to the anode

Applied Potential % O2 Start % O2 End Current

1.0 V 20.6 % 20.5 % 12 mA

1.1 V 20.6 % 20.4 / 20.5 % 26 mA

1.2 V 20.6 % 20.3 / 20.4 % 40 mA

The final set of MEAs tested were standard water electrolysis MEAs with an Ir/Pt anode

catalyst and a Pt supported on carbon cathode catalyst. On MEA I1 two experiments were run;

first He bubbled through 35% H2O2 was fed to the anode with humidified air fed to the cathode

while applying potentials between 1.0 V and 1.2 V. Then He bubbled through 35% H2O2 was fed

to the cathode with humidified air fed to the anode while potentials between 1.0 V and 1.2 V

were applied. In both experiments no current or change in O2 was observed. However, when the

first above experiment was repeated on MEA I2, with flow rates of 250 ml/minute, a small current

and a small change in O2 were observed, as seen in Table 4.3B.

In both MEA I2 and MEA P3 H2O2 vapor was fed to the anode and small currents were

obtained. This indicates that the Ir/Pt and Pt supported on carbon catalysts are capable of H2O2

electrolysis to O2 as is necessary for our conceptual model to work.

55

Table 4.3 B Test results for MEA I2 with humidified air feed to the cathode and He bubbled through H2O2 feed to the anode

Applied Potential Change in O2 Observed Current Theoretical change in O2 based on current

1.0 V 0.10% 40 mA 0.11%

0.8 V 0% 6 mA 0.02%

0.9 V 0% 20 mA 0.06%

1.5 V 0.30% 149 mA 0.41%

1.2 V 0.15% 81 mA 0.23%

1.1 V 0.10% 53 mA 0.15%

4.4 Carbon Degradation

Towards the end of testing on MEA P1, MEA P2, MEA P3, MEA I1, and MEA I2 there was

a noticeable drop in performance on repeated tests. This is most likely due to degradation of the

carbon supporting the Pt catalyst at the cathode. Zhang et al. (2009) summarizes this type of

carbon degradation by reviewing a number of studies on the degradation of Vulcan XC 72R

carbon supports in Pt supported on carbon fuel cells. They discuss how potentials in excess of 1.0

V degrade the carbon supports through CO2 production as seen in Table 4.4A. Additionally lack

of fuel can speed up the carbon degradation as the applied potential has no pathway other than

the carbon degradation to proceed by. As many of our experiments were unsuccessful in

achieving the desired reaction, the cell was essentially in a state of fuel starvation. Additionally

for the H2O2 electrolysis seen on MEA I2 in Table 4.3B (above), the theoretical H2O2 production

for potentials at or above 1.0 V is higher than the observed. This is consistent with the description

of carbon degradation that we are given by Zhang et al. (2009).

56

Table 4.4 A Surface Carbon Degradation

Cs→ Cs+ + e−

Cs+ + 1/2 H2O → CsO + H+

CsO + H2O → CO2(g) + 2H+ +2e−

* Cs denotes a surface species

At the end of testing MEA C1 and MEA C2, a significant drop in electrochemical activity is

seen, as discussed earlier in section 4.3. This is most likely due to degradation of the Printex L6

carbon catalyst from prolonged high voltage and fuel starved testing as discussed with the carbon

supports above.

57

5. Conclusion and Recommendations

5.1 Conclusion

In this study, oxygen transport through a proton exchange membrane fuel cell was

examined in an effort to develop an efficient and effective oxygen pump. The examined MEAs

were loaded with catalysts such as Pt/C, Printex L6, and PtIrB in varying combinations. Each MEA

was run under a varying number of conditions, including feed stream composition, applied

voltage, and temperature. The change in oxygen concentration of the outlet streams and the

generated current were observed and recorded.

As seen in the results for the vapor electrolysis trials, any significant oxygen generation at

the anode of the cell was unattainable in these experiments. The inability to sustain a large

current during vapor electrolysis inhibits the ability of the cell to transport oxygen across the

membrane. While increasing cell size, stacking multiple membranes, and increasing the cell

operating temperature could increase the oxygen yield, these changes to the cell would not make

it any more viable. The membrane area and stacking number of membranes required to achieve

significant oxygen transport would be too large for the design to be a compact and convenient

size. Increasing the temperature of the cell would inhibit the use of the cell as a safe personal

oxygen generator.

In terms of the two electron reduction transport of oxygen, neither the Pt/C catalyst nor

the Printex L6 catalyst was effective in promoting two electron reduction at the cathode. Oxygen

transport was comparable to that of the electrolysis aided pump, and as such neither system

would be suitable for this application. Based on these results, an attempted scale up of these

systems to achieve an effective full-size model would ultimately prove futile.

58

While the vapor electrolysis aided pump is most likely to be ineffective as a method of

oxygen transport, the two electron oxygen reduction method may show some promise in future

testing. The primary issue with the Pt/C and Printex L6 catalyst is that they failed to foster the

formation of hydrogen peroxide at the cathode of the cell. Further research has revealed a

number of catalysts that are more selective for and much more effective for the reduction of

oxygen to hydrogen peroxide. Any further research on this project must begin with the

examination of different catalysts for the oxygen reduction reaction. Subsequent design of the

device can most likely be successfully continued from that point.

5.2 Recommendations The encouraging preliminary results of this project should be followed up with additional

research as suggested to determine the feasibility of oxygen pumping across ion exchange

membrane for the development of an oxygen generator.

5.2.1 Use of Alternative Catalysts

When considering the use of catalysts to assist in this reaction, the following catalysts come

to mind: Pt/C, Pt/Ni and Pt/Ag. This project was only able to conduct preliminary testing on

carbon catalysts. Furthermore, the team has found related research suggesting a variety of

catalysts to test at the cathode side.

Research studies have shown promising results regarding the electrochemical reduction of

oxygen to hydrogen peroxide will be discussed next. For the PEM oxygen pump design, the

cathode side reaction involves the reduction of oxygen to (or production of) hydrogen peroxide.

It is evident that the reaction will require an active, selective and stable catalyst to catalyze the

reaction. Siahrostmi’s study shows Pt-Hg (mercury) to be promising through initial calculations

59

(Siahrostami et al., 2013). In addition, electrochemical measurements suggest Pt-Hg

nanoparticles shows more than an order of magnitude of improvement in mass activity as seen

in the Figure 6.1A below (Siahrostami et al., 2013). It can be seen from Figure 6.1A that the

activity of only platinum is very inefficient. Therefore, Siahrostami results and this project study

on carbon should be taken in consideration for further experiments.

Figure 6.1 A Overview of different electro catalysts for H2O2 production (Siahrostami et al., 2013)

Another study focusing on the anode side reactions suggests a possible catalyst layer that

could be studied to help control the production of H2O2. This study shows a decrease in cathode

open circuit voltage, OCV, correlates to the amount of H2O2 generated within the membrane

(Jung, 2007). The study confirms that a PEMFC with a Pt/ RuC layer at the anode, experiences a

high OCV thus suggesting a lower concentration of H2O2. Though Jung’s findings show do not

60

directly relate to this section, it important to understand potential relationships between testing

parameters.

5.2.2 Fabricating Membranes

As the alternative catalysts suitable for future tests are to be explored, a recommendation is

made to fabricate MEAs with these catalysts accessible commercially. When considering the

further testing of MEAs fabricating, MEAs in the lab will help ensure proper preparations of

membranes since the preparation procedure will be consistent for each MEA created. A sample

procedure has been provided in Appendix C. Further research describes some common

procedures used to load catalysts onto membranes such as the use of spray gun to apply the

platinum onto the membrane under an infrared lamp (Leimin et al., 2009). Additionally, a sputter

technique has been proven to be a useful method to apply minimal amounts (such as

nanoparticles) of platinum on PEMs especially onto Nafion® (Wee et al., 2010). This method could

potentially be adapted for to construct the Pt/Hg-C nanoparticles membrane previously

discussed. The primary benefit for fabricating membranes by hand is the freedom to test various

catalysts.

61

5.2.3 Stacking

If future results are more promising than those in this study, scale up would require testing

MEAs in series as seen in the Figure 6.3A, which will be important for the scale up requirement

for the oxygen generation device. In order to achieve the desired volumetric flow of enriched O2

multiple MEAs together otherwise, i.e., a fuel cell stack, will need to thoroughly investigate. The

project team suggests focusing on the total current density achieved, amount of O2 produced

and amount of time need to produce O2.

Figure 6.3A Sample Fuel Cell Stack (Fuel Cell Store, 2013) Figure 6.3 A Sample Fuel Cell Stack (Fuel Cell Store, 2013)

62

5.2.4 Anion Exchange Membranes

Figure 6.4 A Sample Anion Exchange Membrane for metal cation-free alkaline fuel cell

PEMs are not the only the type membranes viable for conducting such experiments. Anion

exchange membranes (AEM) behave similar to PEMs, however, instead of the proton (+) a

negative charge (-) passes across the membrane. The restriction on time did not allow for the

testing of AEM membranes, however, background section 2.4 on AEM will be useful for further

studies. Utilizing MEAs based on AEMs, which might prove to be more effective than based on

PEMs.

5.2.5 Mathematical Analysis

Mathematical analysis can be insightful. A useful recommendation for future researchers is

to develop a PEM fuel cell (PEMFC) model for oxygen pumping using COMSOL Multiphysics.

COMSOL Multiphysics is a software package that allows for an interactive environment for

modeling and simulating scientific and engineering problems. The model can be used a tool to

63

better understand the physics of PEMFC based oxygen pump. The following figure is an example

of the PEMFC modeled in COMSOL.

Figure 6.5 A Geometry of a proton exchange membrane (PEM) modeled in COMSOL Multiphysics. (COMSOL Multiphysics, 2014)

The makers of COMSOL have provided the following example of an analysis when modeling a

PEMFC: Ohmic Losses and Temperature Distribution in a Passive PEM Fuel Cell

A sample graph from the results of the case study above can be found in Appendix D. The main

recommendation for future researchers is to develop and use a COMSOL model as a tool to gain

a better understanding on the concept of the oxygen pump and testing parameters. In addition,

COMSOL can help provide a theoretical approach and provide a basis on what to expect before

conducting experimental in the laboratory.

64

Works Cited Arges, C.G., Ramani, V, and Pintauro, P.N. (2010). Anion Exchange Membrane Fuel Cells.

Electrochem. Soc. Interface. 19. 31-35.

Assumpção, M. H. M. T., De Souza, R. F. B., Rascio, D. C., Silva, J. C. M., Calegaro, M. L., Gaubeur,

I., Paixão, T. R. L. C., & Hammer, P. (2012). A comparative study of the electrogeneration of

hydrogen peroxide using vulcan and printex carbon supports. Carbon, 49(8), 2842–2851.

Benziger, J., Kimball, E., Mejia‐Ariza, R., & Kevrekidis, I. (2011). Oxygen mass transport limitations

at the cathode of polymer electrolyte membrane fuel cells. AIChE Journal, 57(9), 2505-2517.

Boehm, H. P., Mair, G., Stoehr, T., deRincon, A. R., & Tereczki, B. (1984). Carbon as a catalyst in

oxidation reactions and hydrogen halide elimination reactions. Fuel, 63(8), 1061-1063.

Brillas, E., A. Maestro, M. Moratalla, and J. Casado. "Electrochemical extraction of oxygen from

air via hydroperoxide ion." Journal of applied electrochemistry 27, no. 1 (1997): 83-92.

Choi, P., Jalani, N. H., Thampan, T. M., and Datta, R., “Consideration of Thermodynamic,

Transport, and Mechanical Properties in the Design of Polymer Electrolyte Membranes for Higher

Temperature Fuel Cell Operation,” J. Polymer Sci. B: Polymer Phys., 44, 2183 - 2200 (2006).

Coms, F. D. (2008). The chemistry of fuel cell membrane chemical degradation.ECS

Transactions, 16(2), 235-255.

COMSOL Blog. (n.d.). PEM Fuel Cell Modeling Examples. Retrieved April 23, 2014, from

http://www.comsol.com/blogs/pem-fuel-cell-modeling-examples/

Eladeb, Boulbaba, Caroline Bonnet, Eric Favre, and François Lapicque. "Electrochemical

extraction of oxygen using PEM electrolysis technology." Platinized titanium dioxide electrodes

for methanol oxidation and photo-oxidation: 211 (2012).

Fox, E. B., & Colón-Mercado, H. R. (2011). Mass transport limitations in proton exchange

membrane fuel cells and electrolyzers. Mass Transfer—Advanced Aspects.

Fuel Cell Store. (n.d.). FuelCellStore.com. Retrieved April 23, 2014, from

http://www.fuelcellstore.com/en/pc/viewcategories.asp?idCategory=8

Fujita, Y., Nakamura, H., & Muto, T. (1985). An electrochemical oxygen separator using an ion-

exchange membrane as the electrolyte. Journal of Applied Electrochemistry, 16(6), 935-940. doi:

10.1007/BF01006541

65

Greenway, S. D., Fox, E. B., & Ekechukwu, A. A. (2009). Proton exchange membrane (PEM)

electrolyzer operation under anode liquid and cathode vapor feed configurations. International

journal of hydrogen energy, 34(16), 6603-6608.

Grew, K. N., Chu, D., and Chiu, W. K. S., “Ionic Equilibrium and Transport in the Alkaline Anion

Exchange Membrane,” J. Electrochem. Soc., 157, B1024-B1032 (2010).

Gauthier, W. D., Hendricks, M. J., & Babcock, R. L. U.S. Patent and Trademark Office,

(1980). Oxygen enrichimient system for medical use (US4222750 A).

Hydrogen peroxide. In (1974). D. Considine (Ed.), Chemical and Process Technology Encyclopedia.

New York: McGraw-Hill Book Company.

Jung, U. H., Jeong, S. U., Chun, K., Park, K. T., Lee, H. M., Choi, D. W., & Kim, S. H. (2007). Reduction of hydrogen peroxide production at anode of proton exchange membrane fuel cell under open-circuit conditions using ruthenium–carbon catalyst. Journal of power sources, 170(2), 281-285. Langer, S. H., & Haldeman, R. G. (1964). Electrolytic Separation and Purification of Oxygen from

a Gas Mixture. The Journal of Physical Chemistry, 68(4), 962-963.

Leimin, X., Shijun, L., Lijun, Y., & Zhenxing, L. (2009). Investigation of a Novel Catalyst Coated Membrane Method to Prepare Low‐Platinum‐Loading Membrane Electrode Assemblies for PEMFCs. Fuel Cells, 9(2), 101-105. Litster, S., & McLean, G. (2004). PEM fuel cell electrodes. Journal of Power Sources, 130(1), 61-

76.

Litster, S., Sinton, D., & Djilali, N. (2006). Ex situ visualization of liquid water transport in PEM fuel

cell gas diffusion layers. Journal of Power Sources, 154(1), 95-105.

Ma, C. A., & Yu, W. G. (1995). An electrochemical device for oxygen production avoiding the

generation of hydrogen. Journal of Applied Electrochemistry,26(8), 881-885. doi:

10.1007/BF00683751

Mauritz, K. A., & Moore, R. B. (2004). State of understanding of Nafion®. Chemical

reviews, 104(10), 4535-4586.

Merle, G., Wessling, M., and Nijmeijer, K., “Anion Exchange Membranes for Alkaline Fuel Cells: A

Review,” J. Memb. Sci., 377, 1-35 (2011).

Panizza, M., & Cerisola, G. (2008). Electrochemical generation of h2o2 in low ionic strength media

on gas diffusion cathode fed with air. Electrochimica Acta, 54(2), 876–878. Retrieved from

http://www.sciencedirect.com/science/article/pii/S0013468608009390

66

Pasaogullari, U., & Wang, C. Y. (2004). Liquid water transport in gas diffusion layer of polymer

electrolyte fuel cells. Journal of the Electrochemical Society, 151(3), A399-A406.

Siahrostami, S., Verdaguer-Casadevall, A., Karamad, M., Deiana, D., Malacrida, P., Wickman, B.,

... & Rossmeisl, J. (2013). Enabling direct H2O2 production through rational electrocatalyst

design. Nature materials, 12(12), 1137-1143.

Sigma Aldrich, Nafion® 117. (n.d.). <i>Nafion® 117, thickness 0.007 in.</i>. Retrieved April 30,

2014, from http://www.sigmaaldrich.com/catalog/product/aldrich/274674?lang=en&region=US

Soltani, R. D. C., Rezaee, A., Khataee, A. R., & Godini, H. (2012). Electrochemical generation of

hydrogen peroxide using carbon black-, carbon nanotube-, and carbon black/carbon nanotube-

coated gas-diffusion cathodes: effect of operational parameters and decolorization study.

Research on Chemical Intermediates, 39(9), 4277-4286. doi: 10.1007/s11164-012-0944-8

Song, C., & Zhang, J. (2008). Electrocatalytic oxygen reduction reaction. (pp. 89-134). Springer

London.

Spurgeon, J. M., & Lewis, N. S. (2011). Proton exchange membrane electrolysis sustained by

water vapor. Energy & Environmental Science, 4(8), 2993-2998.

Sudoh, M., Kitaguchi, H., & Koide, K. (1985). Electrochemical production of hydrogen peroxide by

reduction of oxygen. Journal of Chemical Engineering of Japan, 18(5), 409-414.

Sulfuric Acid Molecule. (n.d.). Global Warming Art RSS. Retrieved April 30, 2014, from

http://www.globalwarmingart.com/wiki/File:Sulfuric_Acid_Molecule_VdW_png

Takenaka, H., Torikai, E., Kawami, Y., & Sakai, T. (1982). Extended Abstracts of the 6th Chlor-alkali

Industry Technology Symposium. Electrochem. Soc. Japan, Kyoto. Pg 17.

Tseung, A. C. C., and S. M. Jasem. "An integrated electrochemical-chemical method for the

extraction of O2 from air." Journal of Applied Electrochemistry 11, no. 2 (1981): 209-215.

Varcoe, J., Bruen, L., Chan, N., Flack, N., Girard, D., Handsel, J., Kizewski, J., Murphy, S., Poynton,

S., Slade, R., Waller, T., and Zeng, R., “Anion-Exchange Polymers and Non-Platinum Catalysts for

Alkaline Polymer Electrolyte Membrane Fuel Cell,” Abstract #606, 218th ECS Meeting,

Electrochem. Soc. (2010).

Vega, J. A., Chartier, C., & Mustain, W. E., “Effect of Hydroxide and Carbonate Alkaline Media on

Anion Exchange Membranes,” Journal of Power Sources, 195(21), 7176-7180 (2010).

Wee, J. H., Lee, K. Y., & Kim, S. H. (2007). Fabrication methods for low-Pt-loading electrocatalysts in proton exchange membrane fuel cell systems. Journal of Power Sources, 165(2), 667-677.

67

Wilson, M. S. (1993). U.S. Patent No. 5,211,984. Washington, DC: U.S. Patent and Trademark Office. Winnick, J. (1990). Electrochemical separation of Gases. Advances in Electrochemical Science and

Engineering, 1(3), 205-214.

Zhang, S., Yuan, X., Hin, J., Wang, H., Friedrich, A., & Schulze, M. (2009). A review of platinum-

based catalyst layer degradation in proton exchange membrane fuel cells. Journal of Power

Sources, 194, 588-600.

Zhou, X., Chen, Z., Delgado, F., Brenner, D., & Srivastava, R. (2007). Atomistic simulation of

conduction and diffusion processes in Nafion® polymer electrolyte and experimental validation.

Journal of The Electrochemical Society, 154(1), B82-B87.

68

Appendix A: Results summary table

MEA

Gas Applied

Potential

Change

in O2

Current Humidifier

Temp Misc.

Cathode Anode

MEA

P1 Hydrogen Air Only ran in fuel cell mode. 11 C

MEA

P2

Humidified

He

Dry He

1.6 V 0% 12 mA 11 C

1.7 V 0.10% 53 mA 11 C

1.8 V 0.40% 108 mA 11 C

1.9 V 0.70% 176 mA 11 C

2.0 V 0.70% 203 mA 11 C

Humidified

Air

Dry Air

1.6 V 0.70% 203 mA 11 C

1.3 V 0.10% 40 mA 11 C

1.5 V 0% 53 mA 11 C

1.6 V -0.10% 81 mA 11 C

1.7 V -0.10% 108 mA 11 C

1.8 V -0.10% 122 mA 11 C

1.9 V -0.10% 108 mA 11 C

2.0 V -0.15% 108 mA 11 C

Dry Oxygen

Dry Air

1.2 V 0% 10 mA 11 C

1.4 V 0% 22 mA 11 C

1.6 V -0.10% 67 mA 11 C

69

1.8 V -0.10% 94 mA 11 C

2.0 V 0% 108 mA 11 C

0.8 V 0% 0 mA 11 C

1.0 V 0% 0 mA 11 C

1.2 V 0% 0 mA 11 C

1.4 V 0% 12 mA 11 C

1.8 V 0% 10 mA 11 C

MEA

C1

Humidified

Air

Humidified

Air 0.4 V - 1.0 V

0% for

all runs

0 A for

all runs 11 C

Dry Air Dry Air 0.4 V - 1.0 V 0% for

all runs

0 A for

all runs 11 C

Humidified

O2 Hydrogen

Run in fuel cell mode for 14

minutes 11 C

Humidified

O2 Dry Air 0.4 V - 1.0 V

0% for

all runs

0 A for

all runs 11 C

70

MEA

C2

Humidified

O2 Dry Air

0.4 V -

1.0 V

0% for all

runs

0 A for all

runs 11 C

1.1 V 0% 0 mA 11 C

1.2 V 0% 12 mA 11 C

1.3 V 0% 26 mA 11 C

He & H2O2

Vapor Dry Air

1.0 V 0% 0 mA 11 C

1.2 V 0% 0 mA 11 C

2.0 V -0.70% 108 mA 11 C

2.5 V -4.20% 135 mA 11 C

4.5 V -3.60% 176 mA 11 C

Humidified

Air

He & H2O2

Vapor

1.0 V -

1.2 V

0% for all

runs

0 A for all

runs 11 C

He & H2O2

Vapor

Humidified

Air

1.0 V -

1.2 V

0% for all

runs

0 A for all

runs 11 C

MEA

P3

He & H2O2

Vapor

Humidified

Air 1.0 V -0.10% 12 mA 11 C

1.1 V -0.15% 26 mA 11 C

1.2 V -0.25% 40 mA 11 C

Humidified

Air

He & H2O2

Vapor

1.0 V -

1.2 V

0% for all

runs

0 A for all

runs 11 C

71

MEA

I1

Humidified

Air

Humidified

Air

1.2 V 0% 40 mA 11 C

1.3 V 0.05% 26 mA 11 C

1.4 V 0.10% 40 mA 11 C

1.5 V 0.10% 47 mA 11 C

1.6 V 0.10% 53 mA 11 C

1.7 V 0.10% 60 mA 11 C

1.8 V 0.10% 67 mA 11 C

Humidified

Air

Humidified

Air

1.4 V 0% 40 mA 40 C

1.6 V 0% 40 mA 40 C

1.8 V 0% 53 mA 40 C

Humidified

Air

He & H2O2

Vapor

1.0 V - 1.2

V

0% for all

runs

0 A for all

runs 11 C

He & H2O2

Vapor

Humidified

Air

1.0 V - 1.2

V

0% for all

runs

0 A for all

runs 11 C

Humidified

Air

Humidified

Air

1.8 V 0% 12 mA 11 C

1.7 V 0% 12 mA 11 C Cell

Temp

1.6 V 0% 0 mA 11 C

Humidified

O2 Liquid Water 1.5 V NA 53 mA 11 C

FR = 63

ml/min

FR = 10

ml/min

Positive potential at

anode

Humidified

Air

Humidified

Air 1.4 V 0.20% 53 mA 80 C 60 C

FR = 60

ml/min

FR = 40

ml/min 1.5 V 0.20% 40 mA 80 C 60 C

Positive potential at anode

72

MEA

I1

(Cont.)

Humidified

Air

Humidified

Air 1.4 V -0.15%

12 mA 80 C

60 C

FR = 60

ml/min

FR = 40

ml/min

Positive potential at

cathode

Dry Oxygen Humidified

Air 1.4 V 0.10%

12 mA 80 C

60 C

FR = 40

ml/min

FR = 60

ml/min

Positive potential at

cathode

Humidified

O2 Liquid Water 1.4 V 0%

0 mA 80 C 60 C FR = 60

ml/min

FR = 10

ml/min

Positive potential at

cathode

73

MEA

I2

FR = 250

ml/min

FR = 250

ml/min 1.3 V 0.20% 81 mA 11 C

Positive potential at anode 1.2 V 0.20% 67 mA 11 C ml O2 per Min

Humidified

Air

He & H2O2

Vapor 1.0 V 0.10% 40 mA 11 C 0.27855

FR = 250

ml/min

FR = 250

ml/min 0.8 V 0% 6 mA 11 C 0.04178

Positive potential at anode

0.9 V 0% 20 mA 11 C 0.13927

1.5 V 0.30% 149 mA 11 C 1.037596

1.2 V 0.15% 81 mA 11 C 0.56406

1.1 V 0.10% 53 mA 11 C 0.369078

Humidified

Air

Humidified

Air 1.5 V 0.20% 84 mA 11 C

FR = 235

ml/min

FR = 235

ml/min 1.3 V 0.20% 53 mA 11 C

Positive potential at anode 1.2 V 0.15% 47 mA 11 C

1.0 V 0.10% 20 mA 11 C

Humidified

Air

Humidified

Air 1.5 V 0.20% 94 mA 61 C

FR = 235

ml/min

FR = 235

ml/min 1.5 V 0.30% 94 mA 65 C

Positive potential at anode

1.5 V 0.10% 108 mA 71 C

1.5 V 0.20% 163 mA 75 C

1.5 V 0.20% 163 mA 80 C

74

MEA

I2

.

Humidified

Air

Humidified

Air 1.5 V 0% 0 mA 11 C

FR = 235

ml/min

FR = 235

ml/min 1.3 V 0% 0 mA 11 C

Positive potential at

cathode 1.3 V 0% 0 mA 60 C

Dry Oxygen Humidified

Air 1.3 V 0% 0 mA 11 C

FR = 235

ml/min

FR = 235

ml/min Positive potential at cathode

Dry Oxygen Liquid Water 1.4 V 0% 0 mA 11 C Cell Temp

unheated

FR = 235

ml/min

Positive

potential at

cathode

1.4 V 0% 0 mA 11 C 60 C

Humidified

Air

Humidified

He 1.6 V 0% 0 mA 11 C 60 C

FR = 235

ml/min

FR = 235

ml/min 1.8 V 0% 0 mA 11 C 60 C

Positive potential at

cathode 2.0 V 0% 0 mA 11 C 60 C

Humidified

He

Humidified

He 1.3 V 0% 0 mA 60 C 60 C

FR = 2.4

ml/min

FR = 3.4

ml/min

Positive potential

at cathode

Humidified

He

Humidified

He 1.6 V 0.50% 72 mA 60 C 60 C

FR = 20

ml/min

FR = 17

ml/min 1.7 V 0% 0 mA 60 C 60 C

Positive potential at anode 1.8 V 0% 0 mA 60 C 60 C

75

MEA

I2

1.8 V 0% 0 mA 60 C 60 C

Humidified

He

Humidified

He 1.7 V 0% 0 mA 60 C 60 C

FR = 35

ml/min

FR = 35

ml/min 1.6 V 0% 0 mA 60 C 60 C

Positive potential at anode 1.5 V 0% 0 mA 60 C 60 C

Humidified

He

Humidified

He 1.6 V 0% 0 mA 60 C 60 C

FR = 63

ml/min

FR = 76

ml/min 1.8 V 0% 0 mA 60 C

60 C

Positive potential at anode

Humidified

He

Humidified

He 1.6 V 0% 0 mA 60 C 60 C

FR = 119

ml/min

FR = 146

ml/min 1.8 V 0% 0 mA 60 C

60 C

Positive potential at anode

Humidified

He Liquid Water 1.6 V NA 40 mA 60 C 60 C

FR = 80

ml/min

FR = 10

ml/min 1.8 V NA 67 mA 60 C 60 C

Positive potential at anode

2.0 V NA 81 mA 60 C 60 C

2.2 V NA 94 mA 60 C 60 C

2.5 V NA 108 mA 60 C 60 C

Humidified

O2 Liquid Water 1.5 V 0% 0 mA 60 C

60 C FR = 63

ml/min

FR = 10

ml/min Positive potential at cathode

76

Appendix B: MEA P1 Results Initial Polarization Curve

Initial Power Curve

0.00

0.20

0.40

0.60

0.80

1.00

1.20

0.00 100.00 200.00 300.00 400.00 500.00 600.00

Po

ten

tial

(V

)

Curreny Density (mA/cm^2)

MEA P1 - Initial Polarization

0.00

0.10

0.20

0.30

0.40

0.50

0.60

0.70

0.80

0.90

0.00 100.00 200.00 300.00 400.00 500.00 600.00

Po

wer

(W

)

Current Density (mA/cm^2)

MEA P1 Initial PowerCurve

77

IR Spectroscopy of MEA P1

Before Experimental

IR Specs of MEA P1. The bottom three tests are the control taken of MEA-P1 initial after opened from packaging. The top blue IR spec was taken after the condition of the membrane.

After Experimental

MEA P1 IR post experimental runs. It can be seen there is relatively no activity at all.

78

Appendix C: U.S Patent 5,211,984 for Membrane Catalyst Loading in MEA

Fabrication (Wilson, 1993)

79

80

81

82

Appendix D: Sample plot generated by COMSOL.